Intelligent and energy aware traffic prediction framework for urban transportation

Abstract

The tremendous growth in transport systems and the increase in the number of vehicles over the last decades have created a significant problem in urban areas, namely traffic congestion. Traffic congestion in roads have been a foremost problem in maximum big cities around the globe; especially cities of the developing countries where roads are not well designed as well as traffics on the roads are poorly managed. Traffic congestion increases fuel consumption, causes air pollution. In recent years minimizing the road traffic congestion has been a significant challenge; many researchers have focused on discovering the causes of traffic congestion. Some recent research works have just identified the cause of traffic jam and suggesting an alternate path to avoid traffic congestion. Besides, traffic forecasting requires accurate traffic model which can analyze the actual traffic condition statistically. Intelligent Transport Systems (ITS) are being designed to develop the quality and sustainability of mobility by incorporating data as well as communication technologies with transport engineering. Besides other studies on ITS from the perspective of artificial intelligence (AI) have also been done. ITS depends on a capillary network of sensors which are conveyed over the roads to provide traffic variables like flow, speed, and density. These variables are monitored by administration to approximate traffic dynamics and apply control operations. This thesis proposes a smart framework for the domain of transportation that performs traffic prediction with fuel consumption model and analyzes the traffic flow congestion using genetic and regression model. Also proposes a traffic light controller with a traffic deviation system using the multi-agent system. First, this framework offers smart traffic prediction and congestion avoidance based on the genetic model to reduce fuel consumption and pollution. The model uses Poisson distribution forprediction of vehicle arrivals from recurring size. This model comprises traffic identi